Background Maternal placental syndromes (MPS)—gestational hypertension, pre-eclampsia and placental abruption/infarction—are more prevalent in women with features of the metabolic syndrome (MetSyn). Both MPS and the MetSyn predispose to left ventricular impairment and sympathetic dominance after delivery. Whether this translates into a higher risk of heart failure (HF) and cardiac dysrhythmias is not known.
Objective To determine the risk of new onset of HF and dysrhythmias among women after a prior MPS-affected pregnancy.
Methods A retrospective cohort study was carried out of 1 130 764 individual women with a delivery in Ontario between 1992 and 2009, excluding those with cardiac or thyroid disease 1 year before delivery. The risk of a composite outcome of a hospitalisation for HF or an atrial or ventricular dysrhythmia was compared in women with and without MPS, starting 1 year after delivery.
Results 75 242 individuals (6.7%) experienced a MPS. After a median duration of 7.8 years, the composite outcome occurred in 148 women with MPS (2.54 per 10 000 person-years) and 1062 women without MPS (1.28 per 10 000 person-years) (crude HR=2.00, 95% CI 1.68 to 2.38). The mean age at composite outcome was 37.8 years. The HR was 1.61 (95% CI 1.35 to 1.91) after adjustment for demographic characteristics, diabetes, obesity, dyslipidaemia and drug dependence or tobacco use, as well as coronary artery disease or thyroid disease >1 year after delivery. The adjusted HRs were minimally reduced by further adjusting for chronic hypertension (1.51, 95% CI 1.26 to 1.80) and were higher in women with MPS plus preterm delivery and poor fetal growth (2.42, 95% CI 1.25 to 4.67).
Conclusions Women with MPS are at higher risk of premature HF and dysrhythmias, especially when perinatal morbidity is present.
- Atrial dysrhythmia
- atrial fibrillation
- atrial flutter
- ventricular dysrhythmia
- heart failure
- gestational hypertension
- metabolic syndrome
- deep vein thrombosis
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- Atrial dysrhythmia
- atrial fibrillation
- atrial flutter
- ventricular dysrhythmia
- heart failure
- gestational hypertension
- metabolic syndrome
- deep vein thrombosis
Atrial fibrillation and atrial flutter, the most common dysrhythmias encountered in clinical practice, arise in about 5% of women aged 50–65 years, with a lifetime incidence rate of about 25%.1 They confer a higher risk of chronic heart failure (HF), cardioembolism2 and death, especially in women.3
Gestational hypertension and pre-eclampsia—the onset of hypertension ± proteinuria after 20 weeks' gestation—are associated with abruption or infarction of the placenta. Each may lead to preterm delivery, poor fetal growth and/or intrauterine fetal death.4 ,5 Though multifactorial in aetiology, these maternal placental syndromes (MPS)—gestational hypertension, pre-eclampsia, placental abruption and placental infarction—are worsened by the individual and aggregate features of the metabolic syndrome (MetSyn).6 ,7 Women affected by pre-eclampsia also exhibit higher rates of the MetSyn,8 left ventricular impairment9 and sympathetic dominance10 several months after delivery, all of which are risk factors for atrial and ventricular dysrhythmias.11 ,12
While MPS have been shown to predict the premature onset of coronary artery disease in women,13 the development of HF and dysrhythmias that reflect cardiac dysfunction has not been formally evaluated in this high-risk group. The HAD MPS Study determined the risk of HF and dysrhythmias among women with a prior MPS-affected pregnancy, but in the absence of overt coronary artery or thyroid disease. We also tested the hypothesis that this risk is amplified by the presence of more severe MPS, as well as more MetSyn features.
We completed a population-based retrospective cohort study in the entire province of Ontario, using linked healthcare administrative databases. All residents are enrolled in the Ontario Health Insurance Plan (OHIP), which covers all aspects of antenatal care, including maternal serum screening and ultrasonography, as well as delivery and postnatal care. The HAD MPS Study cohort comprised all women hospitalised for a first obstetrical delivery of a liveborn or stillborn infant after 20 weeks' gestational age, during the period from 1 July 1992 to 31 March 2009. Since we wanted to minimise the possible effects of recurrent MPS or pregnancy effects on cardiac function and since, for the majority of women, this was probably their first live birth, we limited our cohort to the first delivery. An obstetrical delivery was identified using a main patient service code for an ‘obstetrical delivery’. Any subsequent delivery in the same woman was excluded. We also excluded women aged <14 years or >50 years at the time of delivery, and those with an outpatient or inpatient diagnosis of any cardiac or thyroid disease in the period ≤365 days before the index obstetrical delivery date, as listed in online appendix 1.
Exposures and outcomes
We evaluated the presence of MPS, the exposure of interest, during the index delivery hospitalisation. MPS was defined in the index delivery hospitalisation as any diagnosis of gestational hypertension, pre-eclampsia, placental abruption and/or placental infarction, using the ICD-9 coding system up to 31 March 2002 and ICD-10-CA thereafter (online appendix 1).
The primary study outcome was a composite of hospitalisation for HF or an atrial or ventricular dysrhythmia, again, using ICD-9 and ICD-10-CA coding systems. To minimise the immediate effect of the index pregnancy or delivery on the risk of HF or dysrhythmia, assessment of study outcomes (ie, ‘time zero’) began >365 days after the index obstetrical delivery date. Thus, assessment of study outcomes began on 1 July 1993. Secondary study outcomes comprised the individual factors within the primary composite outcome. For the primary study outcome, we also performed a sensitivity analysis in which time zero was set to 30 days postpartum, as described below.
All hospitalisations and procedures were identified using the Canadian Institute for Health Information Discharge Abstract Database (DAD). This was used to characterise the study exclusion criteria, the index delivery hospitalisation, study outcomes and covariates (online appendix 1). The DAD contains the unique encrypted healthcare number, age and sex of the participant, date of admission and up to 16 diagnoses coded by ICD-9 and ICD-10-CA. Some study exclusion criteria and covariates were also identified in the OHIP database (online appendix 1). The OHIP database contains records of all doctors billing information for outpatient and inpatient services, including a service date and a single diagnosis.
Maternal mortality was retrieved from the Registered Persons database, which contains demographic information and encrypted healthcare numbers for all individuals eligible for OHIP. Income quintile and rurality were defined according to postal code using Statistics Canada census data.
Time-to-event analyses were conducted using Cox proportional hazards regression models, to derive an incidence rate and HR and 95% CI for each study outcome, comparing women with and without MPS (the referent). Time-to-event curves and their respective p values for the composite and individual outcomes were generated from the adjusted Cox models.
For each end-point analysis, a participant was censored (ie, determined not to have had a study outcome event) at the time at which he/she either reached the end of the period 31 March 2009 without having experienced that study end-point event or having died during the period of study. Those who might have emigrated from the province after the index delivery, which accounts for about 0.13% of the entire population, were considered lost to follow-up and treated as being event-free up to 31 March 2009.
In the main model, the HR were adjusted for participant age, length of stay in hospital, socioeconomic quintile, rural residence and multiple gestation at the time of index delivery hospital discharge; diabetes mellitus, obesity, dyslipidaemia, drug dependence or smoking ≤365 days before the index obstetrical delivery date and/or within the index obstetrical delivery hospitalisation; and coronary artery disease, coronary revascularisation or thyroid disease, diagnosed >365 days after the index obstetrical delivery date and up to and including the date of diagnosis of a study outcome event. These co-morbidities and potential confounders were abstracted from the DAD (for hospitalised events) and the OHIP database (for outpatient events) (online appendix 1). Information about parity, prior pregnancy loss and medication use were not available.
In the first sensitivity analysis (sensitivity analysis A), we limited the primary composite outcome to a diagnosis comprising the most responsible diagnosis for hospitalisation—namely, that which most strongly influences the patient's length of stay. Additional analyses explored the risk of the composite outcome in association with each MPS component; MPS and the concomitant presence of preterm delivery <37 weeks' gestation, poor fetal growth and/or intrauterine fetal death; severe pre-eclampsia; and pre-eclampsia plus preterm delivery. Relative to women with neither trait, we also determined the risk of the primary composite outcome in the presence of MPS and pre-existing hypertension, as well as MPS plus one or two or more features of the MetSyn—namely, obesity, chronic hypertension, diabetes mellitus or dyslipidaemia. We did not require formal diagnostic criteria for the MetSyn to be met; rather, we considered each individual feature as relevant in the additive risk of the study outcome. As a second sensitivity analysis (sensitivity analysis B), the primary composite outcome was assessed starting 30 days after the delivery date, with adjustment for coronary artery disease, coronary revascularisation or thyroid disease also starting 30 days postpartum. As a third sensitivity analysis (sensitivity analysis C), we evaluated the risk of the primary study outcome of HF or an atrial or ventricular dysrhythmia in association with MPS and adjusted for the variables included in the main model above and also for maternal hypertension ≤365 days preceding or >365 days after the index delivery hospitalisation.
All p values were two-sided, at a significance level of 0.05. All statistical analyses were performed using SAS for UNIX, V.8.02 (SAS Institute).
The three healthcare databases were linked anonymously using encrypted individual health card numbers. The study was approved by the ethics review board of Sunnybrook and Women's College Health Sciences Centre.
We identified 2 017 304 obstetrical deliveries arising in the study period and excluded 886 540 for the following reasons: the delivery was the second event in the study period (n=766 566), no valid health card number (n=9275), non-resident of Ontario (n=1723), age <14 or >50 years (n=358), coding error (n=29), death within 365 days after the obstetrical delivery date (n=216), or diagnosed with any exclusion-based disease within 365 days before the index obstetrical delivery date (n=108 373), as listed in online appendix 1.
Of the final cohort of 1 130 764 women, 75 242 (6.7%) had a MPS-related diagnosis in the index delivery hospitalisation (table 1). The mean age of those with and without MPS was comparable. A multiple gestation pregnancy was also more common in women with MPS (4.9% vs 1.6%). Of all MPS events, about 42% were gestational hypertension, 35% pre-eclampsia, 15% placental abruption and 12% placental infarction; 3.6% were a combination of either gestational hypertension or pre-eclampsia and another MPS component.
The rates of poor fetal growth, intrauterine fetal death and preterm delivery were higher in women with MPS, as were the rates of chronic hypertension, diabetes mellitus, obesity and dyslipidaemia before delivery and chronic hypertension, coronary and cerebrovascular disease after delivery (table 1).
There were 582 836 and 8 267 931 person-years of follow-up in the MPS and non-MPS groups, respectively, with median (IQR) durations of follow-up of 7.8 (3.5 to 12.0) and 7.8 (3.5 to 12.3) years. The primary composite outcome occurred in 148 women with MPS (2.54 per 10 000 person-years) and 1062 women without MPS (1.28 per 10 000 person-years), equivalent to a crude HR of 2.00 (95% CI 1.68 to 2.38) and an adjusted HR of 1.61 (95% CI 1.35 to 1.91) (figure 1A and table 2). The mean (SD) age at the time of a composite event was 37.8 (7.4) years and the maximum age was 60.5 years. In sensitivity analysis A, restricting the primary composite outcome event to a most responsible diagnosis of HF or dysrhythmia, the crude (2.41, 95% CI 1.90 to 3.06) and adjusted (1.93, 95% CI 1.52 to 2.46) HRs were slightly higher. In sensitivity analysis B, in which the primary composite outcome was assessed starting 30 days after the delivery date, the crude HR was 2.01 (95% CI 1.69 to 2.39) and the adjusted HR was 1.62 (95% CI 1.36 to 1.92). In sensitivity analysis C, in which the main model also included a diagnosis of chronic hypertension before or after delivery, the adjusted HR for the association between a MPS and the composite outcome was 1.51 (95% CI 1.26 to 1.80).
The risk of the individual outcomes of HF and atrial and ventricular dysrhythmias are presented in figure 1B–D and in table 2. HF and atrial dysrhythmias were significantly associated with prior MPS in the adjusted models, but ventricular dysrhythmias were not (table 2).
Of the MPS subtypes, the primary composite outcome was significantly more common in women with placental infarction, but not placental abruption (table 3). The adjusted HRs were also significant in women with gestational hypertension and any pre-eclampsia and especially, with severe pre-eclampsia or pre-eclampsia with preterm delivery (table 3).
The main outcome was especially pronounced in women who experienced MPS in combination with preterm delivery and poor fetal growth (adjusted HR=2.42, 95% CI 1.25 to 4.67) (figure 2). The model examining MPS plus intrauterine fetal death was unstable (data not shown). The risk of the primary composite outcome was also notably higher in conjunction with MPS and the MetSyn, especially two or more MetSyn features (figure 2).
The HAD MPS Study, conducted among over 1.13 million individual deliveries, observed a 61% relative increased risk of HF or premature dysrhythmias in women who experienced MPS, independent of recognised coronary artery disease. The risk was potentiated by markers of more serious MPS, such as poor fetal growth, preterm delivery and severe pre-eclampsia and in the co-presence of features of the MetSyn and was minimally lower upon further adjustment for chronic hypertension.
Strengths and weaknesses
We used pre-existing datasets that largely reflect the entire population of women who delivered in Ontario, with a reasonably long-term duration of follow-up. The study exposures and study outcome were identified using ICD-9 and ICD-10 coding systems within comprehensive linked databases. In a large Swedish study of a comparable database birth registry, the positive predictive value was 95% for the presence of normotension and 96% for pre-eclampsia in comparison with direct chart data.14 During our study, the definition of pre-eclampsia remained essentially unchanged.15 Furthermore, in Ontario, blood pressure and proteinuria are measured at every prenatal visit, nearly all women routinely receive antenatal care and 99% deliver in a hospital. In another study from Washington State of 4541 live birth deliveries, use of administrative data records had a positive predictive value of 87% and a negative predictive value of nearly 100% for placental abruption.16 Thus, the majority of MPS events in our study were probably captured and classified satisfactorily.
We only considered an outcome event arising more than 365 days after the index obstetrical delivery, since pulmonary oedema or peripartum cardiomyopathy may be an immediate complication of pre-eclampsia.17 Even so, starting follow-up from 30 days postpartum made no difference. We used a hard end point of hospitalisation for HF or a cardiac dysrhythmia and, when they were defined as the most responsible diagnosis, the effect size was higher. For the study outcome of HF, the DAD administrative dataset used herein was previously verified through random, masked chart audit,18 with 96% of cases meeting the Framingham criteria for HF.19 A study from Alberta, Canada validated the DAD administrative database coding for atrial fibrillation against 966 detailed hospital charts.20 While we considered outcome events serious enough to warrant hospitalisation, we could not delineate whether the dysrhythmias were paroxysmal or persistent, or whether HF arose from diastolic or systolic left ventricular dysfunction.
A diagnosis of chronic hypertension was based on a validated algorithm using diagnostic codes from hospital and doctors' service claims.21 Neither dyslipidaemia nor obesity codes have themselves been validated and physical measures, such as body mass index, were not available herein. Since specific details about tobacco use were also not obtained, it is likely that we missed many women who smoked but who did not receive counselling. Prescription drug use was also not known, including use of antihypertensive agents. Such unmeasured potential confounders may have biased our estimates of the associated risk between MPS and cardiac disease.
We and others previously observed a doubling in the risk of premature coronary, cerebrovascular and peripheral artery disease following MPS.13 This risk was heightened by the concomitant presence of adverse perinatal outcomes and the MetSyn features. Others have observed a higher risk of cardiovascular disease in women with pregnancy loss.22 For example, in the European Prospective Investigation into Cancer and Nutrition cohort in Heidelberg, Germany, comprising 11 518 women, a history of more than three miscarriages was associated with an approximately nine times higher risk of myocardial infarction (adjusted HR=5.06, 95% CI 1.29 to 20.29) and a history of stillbirth was associated with a HR of 3.43 (95% CI 1.53 to 7.72).22 Since ischaemic heart disease is a recognised cause of HF and dysrhythmia, in this study we excluded women with known coronary artery disease before study entry and then adjusted for its presence after time zero. We also controlled for other causes of atrial fibrillation, including thyroid disease and chronic hypertension.
We observed that women with serious forms of pre-eclampsia—severe pre-eclampsia or pre-eclampsia with preterm delivery or poor fetal growth—are at higher risk of cardiac disease. Although we did not measure blood pressure, the data suggest that serious pre-eclampsia may be followed by persistent hypertension after pregnancy and subsequent left ventricular and atrial dysfunction.23
The specific findings of the HAD MPS Study are supported by emerging data suggesting that women with MPS exhibit markers of early cardiac dysfunction. For example, third-trimester concentrations of N-terminal pro-B-type natriuretic peptide are higher in women with pre-eclampsia than in those who are chronically hypertensive or normotensive.24 Women with pre-eclampsia also have higher cardiac troponin I levels in the peripartum period, despite the absence of symptoms of coronary ischaemia.25 Individuals previously affected by pre-eclampsia show signs of cardiac remodelling 3–6 months after delivery, marked by increased left ventricular wall thickness and left atrial dimensions,9 as well as sympathetic dominance.10 In a recent landmark study, 301 women with a history of MPS underwent echocardiography at least 6 months after delivery.26 Of these women, 30 had impaired left ventricular diastolic function; the main risk factor for diastolic dysfunction was the MetSyn (adjusted OR=6.1, 95% CI 2.6 to 14.0). Each individual MetSyn feature was related to diastolic dysfunction.
Recruiting women after a MPS-affected pregnancy on a case–control basis to assess for subclinical myocardial dysfunction is a logical step in testing the mechanisms that might underlie our findings. For example, late gadolinium-enhanced cardiac MRI is a sensitive and specific method to evaluate myocardial fibrosis and left ventricular remodelling.27
The higher risk of premature onset of HF or cardiac dysrhythmias after MPS, especially in conjunction with perinatal morbidity and the MetSyn, provides clinicians with a new opportunity for early disease prevention and expands our approach to postnatal medical care. Recommendations suggest using a history of MPS as an additional early marker of cardiovascular disease risk in women.28 Based on our study findings, this may also be a marker for HF and atrial dysrhythmias. While these outcomes were uncommon after MPS, the mean age of onset was just 38 years and the time-to-event curves diverged with time. Hence, early prevention may avoid the long-term serious consequences of HF and dysrhythmias.2 ,3 Since obstetricians are probably the first to document that a woman had a MPS event, they may have a responsibility to communicate this information to her primary care provider.
One method to prevent cardiac dysrhythmias or heart failure after MPS might be through weight loss and adoption of a healthy lifestyle. This approach confers about a 1.1 mm Hg reduction in systolic blood pressure for each 1 kg reduction in weight.29 A second method might be through pharmacological treatment.30 ,31 Certainly, a variety of antihypertensive agents can be used while breast feeding,32 if needed. A 25% relative risk reduction for new-onset HF was observed in nearly 100 randomised clinical trials of antihypertensive drug treatment in non-pregnant adults.33 Assignment to ACE inhibitors or angiotensin II receptor blockers was also associated with a lower risk of atrial fibrillation (OR=0.65, 95% CI 0.55 to 0.76) among 26 randomised controlled clinical trials.34
Since women with the MetSyn are especially prone to cardiovascular disease,35 earlier identification of those at risk might be enhanced by the addition of MPS to current vascular disease risk models. Prospective cohort studies are now underway to evaluate the 10-year cardiovascular event risk among women with a history of pregnancy-induced hypertension by means of direct assessment of modifiable cardiovascular risk factors.36
Maternal placental syndromes (MPS)—gestational hypertension, pre-eclampsia and placental abruption/infarction—are more prevalent in women with features of the metabolic syndrome (MetSyn).
We found that women with a prior MPS were at significantly higher risk of hospitalisation for heart failure (HF) or an atrial or ventricular dysrhythmia than women without MPS, starting 1 year after delivery.
Women with MPS are at higher risk of premature HF and dysrhythmias, especially when perinatal morbidity is present.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Download Supplementary Data (PDF) - Manuscript file of format pdf
See Editorial, p 1109
This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred.
Funding Funded by the Institute of Human Development, Child and Youth Health, Canadian Institutes of Health Research.
Competing interests None.
Ethics approval Ethics approval was provided by Sunnybrook Health Sciences Centre.
Provenance and peer review Not commissioned; internally peer reviewed.